Say what the page should say.The agent writes it.
Hold the dictation key and talk — the text lands in the agent’s input, recognised on your own machine. The agent reads your style guide from a Reference card and the real code, drafts the page into a note, and shows it on your docs dev server in a browser card.
Two spoken prompts, one webhooks guide. Allow the new file when it asks.
You talk, the agent writes. It reads your style guide and the real code, drafts the page into a note, previews it on the docs dev server — and takes the next correction by voice too.
Talk instead of typing
Hold the hotkey for push-to-talk or tap it to toggle. The text goes into the active agent — you press Enter.
On your machine
Parakeet or Whisper run locally; your voice never leaves the computer.
Grounded in your sources
Style guides and specs on a Reference card, the real code in the repo — no invented header names.
See the real page
The draft is a Markdown note, and the rendered page is in a browser card on your own docs server.
From voice to published page
You speak twice. The agent does the rest.
Good for
Agents for the rest of the work
The same cards run your inbox and your routines. Each page below is a live, clickable scenario.
Eight ways to ship with a squad
Each page is a live, clickable run of a real development workflow — with the same cards and tools the app gives your agents.
Write your next page out loud
Local speech recognition, your own agents.
macOS and Linux are coming later.
The first-run wizard installs an agent CLI, Git and the dictation model for you — or skips each step.